Recording at 24/96K vs. 16/44.1??

[BigRay]

Oh yes, I am aware of that Article, and own the book. Location Classical work is my gig, and yep, some of my favorite recordings were made by Decca and RCA using minimal mic techniques. The reason I said "huh" is because most classical engineers now(the big guys I know anyway) are proponents of massive mic counts and edits. (why a lot of classical recordings now blow)...I thought you were saying that the london philharmonic is recorded now with 2 tracks, which in this multimic world isnt possible..big guys would laugh at that. (myself, I am somewhere in between, I normally use 6-12 tracks, but never any more, as I consider the ensemble to be the most important factor, rather than any one instrument, and dont consider it my place to make balance calls....that is the conductor/composers vision..not mine) I spend HOURS making placement decisions and let the balancing happen naturally..as well as EQ, etc..

Well, how many ears do you have anyway?

I hope you’re aware that some of the best live symphony stereo recordings extant were recorded with a single coincident pair and a half-track analog recorder – just like being there (if you have two ears).

EDIT: Added
Ah, here we go. A good read from Mr. Bartlett himself:
http://www.tape.com/Bartlett_Articles/stereo_microphone_techniques.html

 

[Hollowdan]

Hey werner,
I have tested the theory a little more in depth than any tests I have seen. I will explain how I came to my opinion before I give it, here goes:

I sat a very professional guitarist who I know can play the same thing the same way as many times in a row as needed to the T, in front of a matched pair of mics, with an acoustic guitar. Then I recorded him in 16bit 44.1khz and fast as possible without touching anything, recorded the same song the same way in 24bit 48khz, and again in 24bit 96khz, once more in 24bit 192khz.

BEFORE LISTENING to the different takes in different bit rates!!!
I dithered them all down to 16/44.1 like they would be on the final master.

Then a group of us all blindly listened back to the 4 tracks in 16bit 44.1khz
After listening we also watched them on spectral analyzers.
The verdict was that even blindly everyone in the room agreed that 24 bit 96 khz sounded a million times better than any of the others, yes even the 192khz didn't sound as good as the 96khz take. I am assuming that it had more to do with the dithering not truly being able to handle the 192khz. I have a feeling that 192khz is not truly supported bye anything yet and it will be a couple of years before it becomes actually viable. Also we noticed that indeed the 24/96 song (even after being dithered to 16bit) had obviously lower frequencies and obviously higher frequencies than the lower bit rates. NO JOKE! The 24/48 did NOT go to 40hz and below but the 96khz did!

But I assure everyone here that if you record your audio with good mics and good pre's in 24 bit 96khz and dither it down to 16/44.1 in the mastering phase you will notice a much better sounding recording.
I did NOT want to prove that to be honest because it meant I had to go out and build yet another new computer for my studio lol. I was an unbeliever before that test thinking that this bit rate thing was just a scam for manufacturers to convice the ignorant masses they needed to buy the hottest new product. But alas it is actually true that tracking in 24/96 WILL IN FACT sound better in the final product. I will say that if your computer cannot handle doing ALL your tracks in 24/96 then either get a computer that will or just stay a lower bit rate. Dithering is harmful and should ONLY be done ONCE! I record and mix everything in one bit rate and mix it down in the sme bit rate and do the dithering VERY LAST thing before I actually burn the mastered disk. I even do the mastering in 24/96. I dither it down to 16/44.1 after the song is the way I want it, right before I actually burn the master disk.

I hope this helps and I apologize for letting you know that you will most likely need to upgrade your computer as I had to lol.

One last thing I would like to add is that the quality of the mic's, pre-amp's and compressor's as well as the skill of the person turning the knobs on them, definately effects the outcome of your recordings even more than the sample rate, if you record crap in high quality it is still crap, you can just hear it better.

 

[beezelbubba]

This thread is stupid!
Plug in your instruments and make some music at whatever bit depth/sample rate you like,you friggin' geeks!

 

[Brackish]

Hollowdan, was there any difference noticed between
16/44 and 24/48?

 

[SouthSIDE Glen]

I sat a very professional guitarist who I know can play the same thing the same way as many times in a row as needed to the T, in front of a matched pair of mics, with an acoustic guitar. Then I recorded him in 16bit 44.1khz and fast as possible without touching anything, recorded the same song the same way in 24bit 48khz, and again in 24bit 96khz, once more in 24bit 192khz.

What was your signal chain from beginning to end (mics, preamps, converters, recorders, dithering utility, and anything else you can think of)?

NO JOKE! The 24/48 did NOT go to 40hz and below but the 96khz did!

That you explain how? Where is the weak link in your chain? There is NO technical or even theoretical reason for bit rate resolution to have such an effect on LF response. If there was such a difference, it was most likely an artifact of something in your signal chain or process, not a direct result of the different bit rates themselves.

But I assure everyone here that if you record your audio with good mics and good pre's in 24 bit 96khz and dither it down to 16/44.1 in the mastering phase you will notice a much better sounding recording.

That's a pretty strong assurance based upon the testing procedure used.

A true test would need to be recording the exact same performance to two or three different bit rate recordings simultaneously. Just a 5°-10° shift in the direction of the face of the acoustic guitar between performances could alone account for the difference in LF response you saw and have an equally signifigant difference on the high end. It would also throw a blind testing aspect into the performance end of it, and not just the listening end.

You'd also need to try it with more than one A/D converter. Who's to say that your results aren't simply reporting the results of the performance of that particular converter? Use another converter and maybe the 192k will indeed sound superior, or even both the 96 and the 192 will be worse then the 48. Without having something else to compare it to, you don' know how much of your testing is testing the equipment and how much is testing the sample rate. It's also impossible to tell at this point whether the weakness in the improvement in the 192 response was a reflection of the quality of the rest of the signal chain or of the sample rate.

One last thing I would like to add is that the quality of the mic's, pre-amp's and compressor's as well as the skill of the person turning the knobs on them, definately effects the outcome of your recordings even more than the sample rate, if you record crap in high quality it is still crap, you can just hear it better.

That is the most accurate statement in the post. You did forget to mention the converters, though, which are just as important as the rest of the signal chain, especially when talking about wheteher one can hear the small nuance differences between bit rates.

It may look like I'm slamming your reoprt, Dan. I'm not at all, I appreciate the work you put into it and it's a very helpful post! It does shed light on many things.

All I'm saying that that your "assurance" based upon this one test is a bit premature. Your conclusion is based upon a limited testing procedure that still has too many open-ended variables built into it to make any reliable conclusion.

All you can say for sure is those were the results of single-blind (apparently not double-blind?) listening tests on a single recording chain with no control or baseline to seperate the rest of the chain from the data rate, and that the test data itself (the performance that was recorded) changed from data rate to data rate. No extrapolations or conculsions can truthfully be made beyond that.

G.

 

[NYMorningstar]

Hey werner,

BEFORE LISTENING to the different takes in different bit rates!!!
I dithered them all down to 16/44.1 like they would be on the final master.

If you're doing this all in one step then your results aren't as focused as they could be. You cannot dither a sample rate, you dither a wordlength. If you want results for sample rate then just change the sample rate. I think you will have similar results though and find there is a slight advantage to mixdown using a higher sample rate. Bottom line, work at the highest sample rate and wordlength available, sample rate convert next to last, and dither last.

 

[Hollowdan]

Hollowdan, was there any difference noticed between
16/44 and 24/48?

OH YEAH! HUGE difference!! I would say that the biggest jump was from 24/48 to 24/96 actually which was shocking to us. But definately the second biggest was from 16/44.1 to 24/48. Try it yourself record an acoustic guitar track in 16/44.1 then again the same song in 24/48 with the same signal chain back to back then dither them both down to 16/44.1 and hear for yourself. I know that alot of people are not computer geek's like I have become lately and either cannot, or cannot afford to, build a computer that will do 24/96. It MUST be a computer made specifically for recording, it is insane to see what 24 tracks of 24/96 audio running in full duplex in real time will do to a normal computer! But I can assure you that if you have enough computer to do 24/48 you will see a noticable difference in the sound quality of your recordings. Also I can assure you that if you are recording at 24/96 and are wondering about going to a 64bit windows platform? You WILL once again see a substantial improvment in sound! I kid you not even at the same bit rate, if you upgrade your computer from 32bit to the new 64bit platform your recordings sound much much better! Remember when alesis went from 20bit to 24bit? Remember how much better the unit's sounded? Yeah, it's just like that! Every step technology makes towards giving us more processing power, is one step closer to that pristine sound we all hope to achive in our recordings. I love and hate technology at the same time. I love it because my recordings have never sounded so good (I will NEVER wish for cassette tapes to make a comeback, burn those damn things and let us try and forget they ever existed lol) However I hate technology because it just keeps costing me money! lol. Hope this helps someone, and let's have fun knowing that we have it so much better than the producers of the past! We are truly spoiled compared to the poor guy's of yesteryear.

 

[Hollowdan]

What was your signal chain from beginning to end (mics, preamps, converters, recorders, dithering utility, and anything else you can think of)?

That you explain how? Where is the weak link in your chain? There is NO technical or even theoretical reason for bit rate resolution to have such an effect on LF response. If there was such a difference, it was most likely an artifact of something in your signal chain or process, not a direct result of the different bit rates themselves.

That's a pretty strong assurance based upon the testing procedure used.

A true test would need to be recording the exact same performance to two or three different bit rate recordings simultaneously. Just a 5°-10° shift in the direction of the face of the acoustic guitar between performances could alone account for the difference in LF response you saw and have an equally signifigant difference on the high end. It would also throw a blind testing aspect into the performance end of it, and not just the listening end.

You'd also need to try it with more than one A/D converter. Who's to say that your results aren't simply reporting the results of the performance of that particular converter? Use another converter and maybe the 192k will indeed sound superior, or even both the 96 and the 192 will be worse then the 48. Without having something else to compare it to, you don' know how much of your testing is testing the equipment and how much is testing the sample rate. It's also impossible to tell at this point whether the weakness in the improvement in the 192 response was a reflection of the quality of the rest of the signal chain or of the sample rate.

That is the most accurate statement in the post. You did forget to mention the converters, though, which are just as important as the rest of the signal chain, especially when talking about wheteher one can hear the small nuance differences between bit rates.

It may look like I'm slamming your reoprt, Dan. I'm not at all, I appreciate the work you put into it and it's a very helpful post! It does shed light on many things.

All I'm saying that that your "assurance" based upon this one test is a bit premature. Your conclusion is based upon a limited testing procedure that still has too many open-ended variables built into it to make any reliable conclusion.

All you can say for sure is those were the results of single-blind (apparently not double-blind?) listening tests on a single recording chain with no control or baseline to seperate the rest of the chain from the data rate, and that the test data itself (the performance that was recorded) changed from data rate to data rate. No extrapolations or conculsions can truthfully be made beyond that.

G.

WOW It's a good thing I can take constructive critisism well lol.

You are right maybe I did not give enough information in my first post (or the one I just now posted before reading this one for that matter. I hope this is the last question for me I am tired and should get to bed soon lol)

Signal chain was as follows:
1964 gibson acoustic jumbo (with as I mentioned a VERY professional guitarist who definately played the song as close to identical as humanly possible each pass) Yes I DO know that human's are failable, but he had our faith for this test and I feel he did an excellent job. We simply cannot afford 4 computers to do the test the "correct" way. Nor do we think that our tests were possibly "in-correct" we are very confident in our abilities to perform the test's well enough to prove or disprove what we set out to find out about.

Matched pair of Samson co2's in xy pattern at the 12th fret panned to the sides(I cannot stress enough how good Samson condensor mic's are for the money)
with a Nuemann tlm193 centered. (yep cheap mics AND awesome mics to make sure it wasn't one sided in the respect of only using high end mic's)

zaolla cables
Manley mic pre's
zaolla cables
Apogee rosetta converter's
monster spdif cable
EMU 1820M (mastering grade 24bit 192khz)
Pentium 4 extreme custom built recording computer
Cakewalk sonor 4 (producer edition)
Dithering and spectral analyzing done bye Steinberg's Wavelab

I stated in the first post that I thought the 192khz test was incomplete because I do not think that it is fully supported bye wavelab yet. It did sound as good as the 24/96 but to us at least it did not sound any better. Therefore we determined that it was not yet worth the extra expense to go that route.

For the record:
We did this test for OURSELVES! Only in an attempt to DISPROVE the bit rate hoopla! We WANTED the tests to FAIL because we did NOT want to spend more money on our studios. We are all starving musicians with day jobs that record music simply out of love for music. We whole heartedly wanted to prove to ourselves that the bit rate thing was a scam. However we have all built much better computers, spent alot more money, and bought ALOT more gear as a result of actually hearing the difference with our own ears. We did not do this testing to have it published, or to persuade anyone else into upgrading their gear. We simply did the test for ourselves. I stated my findings on here because I thought it may help someone from spending their hard earned money on stuff they will only want to upgrade in the near future. I hate to tell everyone but believe me or not, that is your choice, and I will hold no ill will against anyone who does not believe me, but everyone will soon discover that as much as we would like to believe that the companys invented the bit rate thing to make us buy more stuff, they actually are telling the truth. 24/96 Does make your recordings cleaner, more open, more present, and all around better sounding.

I truly believe that all the "open ended variables" in the world could not make 16/44.1 or 24/48 sound even close to 24/96 let alone better. The results were atronomical bye sound quality terms. I have recorded many many things the past 16 years I have been into recording, and I personally have enough faith in myself and the 2 other producers in the room to be able to take into consideration such "open ended variables" I have changed mic positions many times and not heard as significant difference in the actual "sound quality" as I did between the different bit rates. We were not listening to the performance, we were listening to the overall SOUND of the recording.

We did also use techniques like actually measuring and marking every inch of the guitar, person, sitting position, and microphones. In an attempt to cut down on those "open ended variables" I kinda like repeating that, everybody now "OPEN ENDED VARIABLES!!!" lol.

With all that said, as far as a "true test"? Personally I have proven all I needed to prove to myself and my friends to justify spending alot of my hard earned money on upgrading my studio to 24/96. You will have to look elsewhere for that "true test" if you seek it. I can only tell you what I experienced, and offer you my opinion on the matter. What is best for you is up to you and no-one else.

OK I hope that covers it, because its my bed time lol. Goodnight all!

 

[Hollowdan]

If you're doing this all in one step then your results aren't as focused as they could be. You cannot dither a sample rate, you dither a wordlength. If you want results for sample rate then just change the sample rate. I think you will have similar results though and find there is a slight advantage to mixdown using a higher sample rate. Bottom line, work at the highest sample rate and wordlength available, sample rate convert next to last, and dither last.

Ummm......"what?" out of curiosity are you just trying to correct me for no other reason than to correct me? I honestly do not see a point to your post, maybe I am just tired, I will attempt to read your post again tomarrow and see if I can actually derive a statment/question/response from it and reply properly.

Thank you for taking the time to read my post and reply, just not sure what your reply is getting at, have a good night.

 

[NYMorningstar]

Ummm......"what?" out of curiosity are you just trying to correct me for no other reason than to correct me? I honestly do not see a point to your post, maybe I am just tired.

Sorry for the confusion. I was just suggesting that if you want to test the differences between wordlengths, then you should only change the word lengths, e.g. go from 44100 hz/32 bit to 44100 hz/16, 44100/24 to 44100/16 or 44100/20 to 44100/16. That way you'd just be testing the difference in sound quality of only the change in wordlength and not the combination of wordlength and sample rate.

 

[SouthSIDE Glen]

Sorry for the confusion. I was just suggesting that if you want to test the differences between wordlengths, then you should only change the word lengths, e.g. go from 44100 hz/32 bit to 44100 hz/16, 44100/24 to 44100/16 or 44100/20 to 44100/16. That way you'd just be testing the difference in sound quality of only the change in wordlength and not the combination of wordlength and sample rate.

Moriningstar is absolutely correct.

Dan, I wasn't criticizing the abilities of you or your cohorts. All I was saying was that the design of the testing you did was such whereas it's inpossible to reach the conclusions you did with any degree of certainty because the actual design of the test was flawed, no matter how well you executed it.

I'm not just picking nits here. The fact reamins that all you can state with certainty would be "here were the subjective results obtained with this signal chain under these conditions." You have nothing to test those results against.

Were the results you heard due to the difference in sample rates or due to the performace characteristics of the Apogee or of Wavelab's out-of-the-box dithering and rate conversion? There's absolutely no way to provide an answer to that from this test, because those were the only devices used and because there was no control sample to compare the results against.

How much and what kind of a difference was caused by the sample rate conversion vs. the bit depth dithering? Again, apparently we can't tell because they were both performed at the same time. Based upon the test structure itself, it's impossible to tell where the difference were due to the sample rate or due to the characteristics of the dithering algorithm.

And none of this yetl explains where that low frequency content came from. Your attribution to the extra LF coming from the higher sample rate is almost certainly false, based simply on the laws physics and mathematics. So if the LF didn't come from the thing we were testing for - i.e. the difference in sample rates - then it came from somewhere else. The problem is that if one wants to test something, all the "somewhere elses" have to be locked down tight, otherwise the test results are meaningless. The fact that you have a large, unexplained bump in LF in the 96k recording most likely indicates a diiference in the source recording. I'm sure you were very careful, but no matter how careful one is, one cannot deny the results or the fact that they cannot be readily explained as a positive result for what you were testing for.

If after all this, it's still not good enough to convince you, I'll not try to talk you out of it. I'm sure you're going to incorporate this into a sales pitch to sell some poor schmuck who can barely play an instrument on upgrading the working bandwidth of his system because it "sounds better" (you can't tell a customer the whole GIGO angle of it because you can't chance insulting the customer.)

But please don't try to lay that boogie woogie here. Your test did what it did and it did it well. But by it's own design it simply cannot make any statements about the intrinsic differences between sample rates.

G.

 

[Hollowdan]

Moriningstar is absolutely correct.

Dan, I wasn't criticizing the abilities of you or your cohorts. All I was saying was that the design of the testing you did was such whereas it's inpossible to reach the conclusions you did with any degree of certainty because the actual design of the test was flawed, no matter how well you executed it.

I'm not just picking nits here. The fact reamins that all you can state with certainty would be "here were the subjective results obtained with this signal chain under these conditions." You have nothing to test those results against.

Were the results you heard due to the difference in sample rates or due to the performace characteristics of the Apogee or of Wavelab's out-of-the-box dithering and rate conversion? There's absolutely no way to provide an answer to that from this test, because those were the only devices used and because there was no control sample to compare the results against.

How much and what kind of a difference was caused by the sample rate conversion vs. the bit depth dithering? Again, apparently we can't tell because they were both performed at the same time. Based upon the test structure itself, it's impossible to tell where the difference were due to the sample rate or due to the characteristics of the dithering algorithm.

And none of this yetl explains where that low frequency content came from. Your attribution to the extra LF coming from the higher sample rate is almost certainly false, based simply on the laws physics and mathematics. So if the LF didn't come from the thing we were testing for - i.e. the difference in sample rates - then it came from somewhere else. The problem is that if one wants to test something, all the "somewhere elses" have to be locked down tight, otherwise the test results are meaningless. The fact that you have a large, unexplained bump in LF in the 96k recording most likely indicates a diiference in the source recording. I'm sure you were very careful, but no matter how careful one is, one cannot deny the results or the fact that they cannot be readily explained as a positive result for what you were testing for.

If after all this, it's still not good enough to convince you, I'll not try to talk you out of it. I'm sure you're going to incorporate this into a sales pitch to sell some poor schmuck who can barely play an instrument on upgrading the working bandwidth of his system because it "sounds better" (you can't tell a customer the whole GIGO angle of it because you can't chance insulting the customer.)

But please don't try to lay that boogie woogie here. Your test did what it did and it did it well. But by it's own design it simply cannot make any statements about the intrinsic differences between sample rates.

G.

To be honest I feel that this post is just down right rude and uncalled for. I have said it MORE THAN ONCE, this test was done for my OWN amusement. I HATE sales guys and if you knew me you would understand that just because I work in a music store does not mean that I am just out to get some "poor shmuck's" money. Besides what do I have to gain bye "trying to sell my boogie woogie" in HERE? I will get NOTHING from it. So in light of that I would prefer not to be insulted bye you for trying to help someone in making a decision. I think this has to be the most close minded and self glorifying post that I have personally ever read.
For the LAST time, "I" can definately tell a very audible difference in tracking at a higher bit rate/word length. You can believe me or not, I am not here to try and sell anyone anything. The post asked people what they knew on the subject and this is what I know. Everyone who has heard the CD we made of the results has agreed with us that the 24/96 recording sounds far better than the others. It has nothing to do with performance, it has to do with how much more present, open, dynamic, and overall warmer it actually sounds. If you would do a little bit of research on the subject you would find that many top name producers will say the same thing "recording at higher bit rates/word lengths simply sounds better" I have even talked to Malcom Toft in person about it and he agree's with me. So trust me I will be able to sleep like a baby tonight wether you agree with me or not. I love the sound of my recordings since I have upgraded to 24/96 and to me that is really all that matters. My recordings sound better now, once again I say "Thank you technology" period.

P.S. this is a forum of people exchanging information on a "friendly" level. To tell another person that you have never met or heard any of their recordings that they are "full of it" and just trying to "sell poor schmucks gear they can't use" in my opinion is very much missing the point. I take everyone's opinion in here as if they know more than me. I TRY everything people tell me on here before I dismiss it. You can learn something from everyone, no-one knows everything, and no-one should feel that they are better than anyone else. Recording is about being an artist, experimenting to find out what works best for "YOU" what may work great for you may not work great for me, and vice versa.

In the time it took you to pick apart my posts and try your best effort to discredit me, you in fact could have TRIED this and found out for yourself.

Do the test in the way you consider "the right way" since my way is apparently BS. Or are you afraid that I may be right?

I would like to apologize to everyone who has to read through this little dispute that is obviously doing no-one any good. I just don't like being insulted and felt that I had to defend myself.

 

[NYMorningstar]

Hey Hollowdan chill. It' your own post - "Ummm......"what?" out of curiosity are you just trying to correct me for no other reason than to correct me?" - that started the rudeness in here if you ask me. SouthSIDE Glen is just pointing out some inacurracies which I'm not sure you even get yet.

Now you're talking crap like "recording at higher bit rates/word lengths simply sounds better". You don't record at higher bit rates. I'm sure you mean sample rates but that's just the kinda stuff you need to be critiqued on if you're ever going to learn this shit. SouthSIDE Glen isn't trying to discredit you, he is just giving what your saying a critique so you don't use it to influence people and spread more inaccurate bullshit around.

 

[SouthSIDE Glen]

Oh, man, there has been a real run of situations on this board in the past week where people take personal insult at posts pointing out factual truths. Is the the heat?

OK, first off I do want to apologize for and retract the section where I talk about the sales pitch. That was premature and uncalled for on my part, and I sincerely paologize for that one. I was wrong there. The heat gets to me sometimes too .

As for the rest of the post, though, I stand by it as am impersonal analysis of the situation. I have no doubt of what you heard. And I'm not telling you that what you heard was wrong or that it was wrong to report it to us here. Please undertand this: what I am saying is that because of the way the test was setup, the conclusions you are drawing from what you heard are premature and unsupported. That's not a slam or an insult...or at least that's not how I intend it. I'm just saying that's the test as it was designed can only tell anybody just so much and that it in fact contained results that indicate just how incapable the thst criteria were to allow anybody to draw a solid conclusion about the intrinsic value of sample rates and bit depths. There is no way to seperate the effect of bit rates from the effect of gear coloration in that test.

That's NOT an insult, for God's sake. That's just pointing out a technical reality. What do you think, that I'm trying to slam you for not being a perfect experimental scientist? Do you really think I'm that petty?

This is a forum of an exchange of ideas and opinions, yes. But while all opinions are created equal, all truths are not. There is a difference between someone who says "I believe this to be true" and "This is the truth". The first is fine. The second is fine only as long as it is indeed true. All I'm saying is that the test allows you (or anybody else) to say the first, but by it's very nature it is incapable of allowing anyone to assert the second.

You "assured" readers that X is true about bit rates and provided the test to back up that assurance. I just felt obligated to point out why that "assurance" should be regarded more as an opinion than as a demonstrated truth because the test was not robust enough to settle the matter.

And to flip sides and join you for a minute here, Dan, (I don't play favorites ) there is nothing technically wrong with using the term "bit rate". I use it all the time. Mshilarious, all "bit rate" means is the number of bits that cross one's path within a certain amount of time. Higher sample rates require higher bit rates - they *are* higher bit rates. Now, I'll admit that the full and specific bit rate in this case would be defined as the sample rate multiplied by the bit depth/word length. But in general terms it is perfectly Ok on a general level to say that (for example) a 48k sample rate is a higher "bit rate" than 44.1k is.

Time for an ice cold Oolog iced tea...

G.

 

[NYMorningstar]

there is nothing technically wrong with using the term "bit rate". I use it all the time. Mshilarious, all "bit rate" means is the number of bits that cross one's path within a certain amount of time. Higher sample rates require higher bit rates - they *are* higher bit rates. Now, I'll admit that the full and specific bit rate in this case would be defined as the sample rate multiplied by the bit depth/word length. But in general terms it is perfectly Ok on a general level to say that (for example) a 48k sample rate is a higher "bit rate" than 44.1k is.

I see where you are going with this but I think it's misleading. While often referred to as "speed", bit rate does not measure distance/time but quantity/time, and thus should be distinguished from the "propagation speed" (which depends on the transmission medium and has the usual physical meaning).


The bit rate depends on several factors:

The original material may be sampled at different frequencies.
The samples may use different numbers of bits.
The data may be encoded by different schemes.
The information may be digitally compressed by different algorithms or to different degrees.

I'd say in general on the audio level that bit rate is used in reference to an MP3. For example:

4 kbit/s — minimum necessary for recognizable speech (using special-purpose speech codecs)
8 kbit/s — telephone quality
32 kbit/s — MW (AM) quality
96 kbit/s — FM quality
128 kbit/s — Typical "acceptable" music quality
256 - 320 kbit/s — Near audio CD quality

The formal abbreviation for "bit per second" is "bit/s" (not "samples/s"). In less formal contexts the abbreviations "b/s" or "bps" are often used, though this risks confusion with "bytes per second" ("B/s", "Bps"). Even less formally, it is common to drop the "per second", and simply refer to "a 128 kilobit audio stream" or "a 100 megabit network".

Referring to a sample rate as a bit rate is doing nothing more than darkening the already muddy waters.IMHO Please tell me if I'm incorrect also because I am just learning this stuff myself.

You were right about the heat. I spent the day in the sun pulling out weeds, bah and more bah. As soon as the sun goes down I'm going out for nice long relaxing run Cheers

 

[SouthSIDE Glen]

Referring to a sample rate as a bit rate is doing nothing more than darkening the already muddy waters.IMHO Please tell me if I'm incorrect also because I am just learning this stuff myself.

You are correct as I see it, and it's correct to say to using "bit rate" as a "generic" term can definitely muddy up the specific technical definition of "bit rate" in any specific situation.

Technically, the "nominal bit rate" - or more appropriately the "data rate" - of a particular recording resolution is a multiple of the sample rate and word length:

96/24 = 2.304 Mbits/sec
48/24 = 1.152 Mbits/sec
44.1/24 = 1.058 Mbits/sec
44.1/16 = 0.705 Mbits/sec (or ~705 kbits, if you want to compare that to MP3)

The actual "throughput rate" of actual bits through the hardware can vary from these figures for several technical reasons, but these are the base theoretical bit rates for each resolution.

In this view, calling just the sample rate alone the "bit rate" is indeed technically incorrect, as the sample rate is only half of the equation.

But it is also a fairly common, if not entirely correct, technical slang to refer to any value that's ultimately based upon a bits per second measurement as a "bit rate". In this case sample rate is a form of "bit rate" because it is measuring the number of samples per second. On the other hand, the bit depth/word length is not a "bit rate" because it is not a time-based measurement.

Think of it this way, the bit depth - 16 or 24 bit - is defining the size of a "byte" of digital audio sample. 44.1/16 has a data rate that can be referred to as 705 kbits/sec, or as 44.1 kbytes/sec with a byte of audio data in this case defined as 16 bits.

Yeah, I guess the language can get confusing. It'd be a lot easier if we all just stuck to the term "sample rate" and "bit depth". Frankly, some of use are used to talking in a shorthand that loosely referrs to any digital "rate" generically as a "bit rate". Maybe we shouldn't. I'll try to break myself of that habit here. But on that same token, I can't rightly fault Dan for using that term either.

G.

 

[Hollowdan]

Hey Hollowdan chill. It' your own post - "Ummm......"what?" out of curiosity are you just trying to correct me for no other reason than to correct me?" - that started the rudeness in here if you ask me. SouthSIDE Glen is just pointing out some inacurracies which I'm not sure you even get yet.

Now you're talking crap like "recording at higher bit rates/word lengths simply sounds better". You don't record at higher bit rates. I'm sure you mean sample rates but that's just the kinda stuff you need to be critiqued on if you're ever going to learn this shit. SouthSIDE Glen isn't trying to discredit you, he is just giving what your saying a critique so you don't use it to influence people and spread more inaccurate bullshit around.

\

Yeah... don't think I was talking to you but thanx for your .02 anyways. As for "inacurate bs" cussboy (kids might read this forum bye the way) I have been trying to explain 2 things that have fallen on deaf ears apparently.

1-I was just trying to answer the guy's question!!!!

2-IF YOU tried this test yourself (which would take much less time than the time it takes to cuss me out on several posts) YOU YOURSELF would come to the same conclusion I have and, believe it or not, agree with me!!!!

So I ask you simply to get off my back, get off your obviously skyscraper high "horse" and get as scientific as you deem required, and TRY IT!

I have ALL the proof I personally need in this matter.
I made a "statment" (I did not state it as a "proven fact") to try and help out a fellow studio dog.
I NEVER intended to be rude, argue, belittle, degrade, mislead, anger, frusterate, or SELL anything to anyone.
You have never heard a recording of mine, nor were you present during the test I refer to, and bye that simple logic I do NOT feel you have the right to judge my "knowledge" of this subject to be "inacurate BS"
Mostly I did not ask, nor imply that I wanted ANYONE'S Critique on anything I said. I feel my recording's are excellent and joined this forum to help other's and maybe pick up on a few idea's I hadn't thought of. I am NOT saying I am better than anyone else, or that my recordings are better than anyone else's. Just simply "stating" that I answered a post. I did NOT ask for anyone's opinion or "critique" on my reply.
As for the "if your going to learn this s@#t" thing: I really feel that you have no right to determine what I "need to learn" I do not "NEED" to be critiqued on anything, Thank you very much.
I have decided that this has gotten so out-of-hand and insulting to the person who posted this originally, that I refuse to come back in here. Cus me out all you feel like, this is my final word on this.

I am willing to bet that one day everyone who has disagreed with my "statments, once again not implied EVER as being proven fact" will realize that 24/96 is in fact a better sounding recording format than 16/44.1. I just wish I could be there to see that last drop of pride as you swallow it lol.


I TRIED to help someone, so shoot me!

 

[Hollowdan]

Oh, man, there has been a real run of situations on this board in the past week where people take personal insult at posts pointing out factual truths. Is the the heat?

OK, first off I do want to apologize for and retract the section where I talk about the sales pitch. That was premature and uncalled for on my part, and I sincerely paologize for that one. I was wrong there. The heat gets to me sometimes too .

As for the rest of the post, though, I stand by it as am impersonal analysis of the situation. I have no doubt of what you heard. And I'm not telling you that what you heard was wrong or that it was wrong to report it to us here. Please undertand this: what I am saying is that because of the way the test was setup, the conclusions you are drawing from what you heard are premature and unsupported. That's not a slam or an insult...or at least that's not how I intend it. I'm just saying that's the test as it was designed can only tell anybody just so much and that it in fact contained results that indicate just how incapable the thst criteria were to allow anybody to draw a solid conclusion about the intrinsic value of sample rates and bit depths. There is no way to seperate the effect of bit rates from the effect of gear coloration in that test.

That's NOT an insult, for God's sake. That's just pointing out a technical reality. What do you think, that I'm trying to slam you for not being a perfect experimental scientist? Do you really think I'm that petty?

This is a forum of an exchange of ideas and opinions, yes. But while all opinions are created equal, all truths are not. There is a difference between someone who says "I believe this to be true" and "This is the truth". The first is fine. The second is fine only as long as it is indeed true. All I'm saying is that the test allows you (or anybody else) to say the first, but by it's very nature it is incapable of allowing anyone to assert the second.

You "assured" readers that X is true about bit rates and provided the test to back up that assurance. I just felt obligated to point out why that "assurance" should be regarded more as an opinion than as a demonstrated truth because the test was not robust enough to settle the matter.

And to flip sides and join you for a minute here, Dan, (I don't play favorites ) there is nothing technically wrong with using the term "bit rate". I use it all the time. Mshilarious, all "bit rate" means is the number of bits that cross one's path within a certain amount of time. Higher sample rates require higher bit rates - they *are* higher bit rates. Now, I'll admit that the full and specific bit rate in this case would be defined as the sample rate multiplied by the bit depth/word length. But in general terms it is perfectly Ok on a general level to say that (for example) a 48k sample rate is a higher "bit rate" than 44.1k is.

Time for an ice cold Oolog iced tea...

G.

Glen I appreciate the apology and would like to apologize myself for taking what you said as insulting. It came across that way and I know things are very easily taken out of context in type. I did not mean to be rude if it came across that way. I have seen several of your replies on here and I know that you are very intellegent and helpful to alot of people. I just felt that I had been attacked needlessly. I offered up a test I did as an answer to help someone with a question they posted and found myself getting a huge backlash telling me that I was "incorrect", "trying to sell people useless stuff", "inacurate", "full of BS" and several other things. I tested the theory of higher sample rates with 2 other of my close friends and trusted engineers, and found more than enough proof to spend my hard earned money to upgrade my studio and have reaped the rewards every since. I tried to explain myself in that, regardless of any possible inaccuracies, we heard enough of a difference in the sound quality that we felt no need to do a more "scientifically perfect" test. We were sold, and considering we are all starving musicians I though that said alot to our findings. I am not going to come into this post anymore because I feel that all the argueing has been disrespectful to the original poster. So once again I apologize, I was just trying to defend myself, and straighten up some misunderstandings.

I never stated that my finding's should be considered religion, or scientific fact.

 

[SouthSIDE Glen]

Let's just forget it happened, Dan. It turned into not the most shining moment for either of us, I suspect .

G.

 

[Farview]

Not to stir anything up, but I just did a comparison test. To take the human element out of it, I reamped a recorded DI guitar signal through the same signal chain twice. At 44.1/24 and 96/24.

The DI recording was played on my DA-88 through on of my old Laneys, miced up with a 421 and an m-149. Both mics were run through my Amek 9098 preamp and into the computer.

I can not tell the difference. There definatly isn't any increased low end, clarity, etc...

I'm trying to figure out how to do this with an instrument that actually has a lot of energy in the top octave, but the repeatability is going to be tough. There is no way to hit a cymbal, for example, and have it sound the same twice. Even if you set up a machine to strike it the same way every time, the two hits will sound different. Acoustic guitar is impossible to keep the position stable, much less the performance.